I've Changed My Mind About AI and Sales Coaching. Here's Why.
real opportunity for artificial intelligence may not be to replace the sales coach, but to help sales managers become better at something we’ve gradually allowed to disappear.
I’ve spent much of my career working with salespeople and sales leaders, and I’ve always believed that coaching is one of the most important responsibilities of anyone managing a sales team. It’s also one of the responsibilities I fear we’ve increasingly neglected.
So when artificial intelligence began appearing in conversations about replacing human sales coaches, my initial reaction was one of considerable scepticism. I wasn’t particularly enthusiastic about the idea of technology taking over something I regarded as fundamentally human.
Coaching, in my experience, is built on relationships. It requires trust, understanding, encouragement, challenge and the ability to recognise when somebody needs guidance rather than another instruction. A good coach doesn’t simply provide answers. They help people develop the confidence and capability to find answers for themselves.
I struggled to see how an AI system could replace that relationship, and frankly, I didn’t want it to.
But I’ve changed my mind. Not about the importance of human coaching, but about the role AI can play in making it more effective.
And the change has come from a better understanding of what the technology can do, combined with our practical experience of introducing AI-supported feedback into structured sales coaching processes.
What I’ve come to recognise is that we may have been approaching the debate from entirely the wrong direction. Rather than asking whether AI will replace human sales coaching, perhaps we should be asking whether it can help us restore and improve a leadership discipline that has been gradually displaced by the pressures of managing sales performance.
Have we become managers of results rather than developers of people?
Over the years, I’ve watched personal development budgets come under pressure, training responsibilities migrate towards digital learning platforms and the completion of online courses become an increasingly convenient measure of development.
There is nothing inherently wrong with digital learning. It can provide valuable knowledge, consistency and accessibility. But watching a video, completing an assessment and receiving a certificate doesn’t necessarily mean that somebody has developed the ability to apply what they’ve learned in a real customer situation.
Meanwhile, sales managers face enormous pressure to deliver results. Forecasts, pipeline reviews, CRM compliance, activity reports and performance dashboards all demand attention. The expectation to deliver this month’s or this quarter’s number is understandable, but the consequence is that management activity can easily take precedence over coaching.
We spend considerable time measuring performance and comparatively little time examining the behaviours that produce it.
A manager can identify an opportunity that hasn’t progressed, challenge the salesperson about the forecast and demand a recovery plan. But none of those activities necessarily develops the salesperson’s ability to understand why the opportunity is stuck or what they might do differently.
This is the distinction that increasingly concerns me. Managing performance and developing performance are not the same thing, although effective sales leadership requires both.
And if we’re not providing regular, meaningful coaching, perhaps the arrival of AI should be welcomed rather than feared. After all, if technology can provide a salesperson with guidance they might otherwise never receive, surely that has some value.
But I now believe there’s a considerably better opportunity than simply giving salespeople access to an AI coach.
Putting the human conversation back at the centre
One of the disciplines we’ve been developing with sales leadership teams is WILO, our structured weekly one-to-one coaching process.
Its premise is straightforward. The sales manager and salesperson meet every week, ideally on the same day and at the same time, with a clear structure and a shared responsibility for preparation.
The conversation is deliberately forward-looking. Rather than spending the meeting explaining what happened last week, the focus is on the business being developed, the opportunities in the pipeline and the decisions and actions required to make progress.
What does the salesperson understand about the customer’s situation? What remains unknown? What evidence supports their assessment of the opportunity? What needs to happen next, and what might prevent it from happening?
Importantly, the manager’s responsibility is not simply to inspect progress or demand activity. It’s to help the salesperson think more effectively, challenge assumptions, explore alternatives and remove barriers that might otherwise slow progress.
Sometimes that will involve sharing experience or offering direct advice. But if every conversation ends with the manager identifying the problem and prescribing the solution, we have to question how much capability is actually being developed.
The objective is to help the salesperson become increasingly effective at making commercial decisions independently, while knowing they have the support of their manager when they need it.
This is particularly important in complex B2B sales, where the salesperson is often dealing with multiple stakeholders, technical considerations, commercial risks and customer decisions that cannot be reduced to a simple sequence of activities.
A well-structured weekly conversation provides an opportunity to develop the quality of the salesperson’s thinking while there is still time to influence the outcome.
And that is where our experience with AI has become particularly interesting.
What happens when we start evaluating the quality of coaching?
Through our recent work with a sales leadership team, we’ve been reviewing recordings and transcripts of weekly coaching conversations against an agreed framework.
The intention was not to evaluate managers simply for the sake of scoring their performance. It was to help them understand what they were doing well, identify opportunities to improve and develop greater consistency in their coaching.
Even in the early stages, the distinction between conducting a meeting and genuinely coaching became apparent.
There were encouraging examples of supportive conversations, commercial understanding and managers helping their people address real customer opportunities. There was clear commitment to making the meetings useful, and the introduction of a consistent weekly structure was itself an important step forward.
But the feedback also highlighted some recurring development opportunities.
Managers could move relatively quickly towards offering solutions rather than allowing the salesperson to explore the problem. Actions could be discussed without sufficient clarity about ownership, timing or evidence of customer commitment. And some conversations were more effective at providing the manager with an update than at helping the salesperson develop their own thinking.
These observations matter because a meeting can be well organised, positive and commercially relevant without necessarily being an effective coaching conversation.
If the manager does most of the thinking, identifies the problem, proposes the solution and determines the next action, the salesperson may leave with a useful plan. But have they developed the judgement and confidence to handle a similar situation independently next time?
That’s a much more demanding question than simply asking whether the weekly meeting took place.
And it’s also where AI can make a meaningful contribution.
Using AI to develop the coach, not replace them
Imagine the weekly conversation continuing exactly as it should, with the salesperson and manager working together on genuine commercial opportunities.
With appropriate consent and safeguards, the conversation is recorded and transcribed. AI then reviews the transcript against an agreed coaching framework, identifying examples of effective questioning, listening, challenge, action planning and the extent to which the salesperson was encouraged to develop their own solutions.
The manager receives feedback on what worked well and what could be improved. Rather than a generic suggestion to ask better questions, the feedback can identify a specific point in the conversation where an alternative question might have encouraged the salesperson to think more deeply.
It can highlight where an assumption went unchallenged, where an action lacked sufficient clarity or where the manager missed an opportunity to explore the customer’s decision-making process.
And importantly, the feedback can establish a specific development focus for the next conversation.
Over successive weeks, the same framework provides an opportunity to examine whether coaching behaviours are changing. Is the manager asking more purposeful questions? Are salespeople contributing more of their own thinking? Are actions becoming clearer? Is the conversation becoming more focused on progressing opportunities rather than reporting activity?
This is the aspect of AI-supported coaching that has fundamentally changed my perspective.
The technology isn’t replacing the relationship between the manager and salesperson. It’s providing structured feedback that can help the manager become more effective within that relationship.
But there’s another important part of our process that I believe deserves particular attention: the AI-generated report is not the final report.
Before any WILO coaching feedback is shared with a sales manager, I personally review and verify it.
That means examining the AI’s observations against the conversation, considering whether the feedback is supported by the evidence and whether the proposed development priorities are fair, relevant and useful. Where necessary, I challenge or refine the assessment before it is issued.
I consider this an essential part of the process rather than an additional administrative step.
AI can analyse a conversation against a framework with considerable consistency, but consistency alone doesn’t guarantee sound judgement. A question that appears ineffective in isolation may make sense within the wider commercial context. A manager who offers direct advice may be doing precisely the right thing at that moment. Equally, a seemingly positive conversation may have missed an important opportunity to challenge the salesperson’s assumptions.
These distinctions matter, particularly when the feedback is intended to influence somebody’s professional development.
The human review provides another level of assurance that the output is not simply technically consistent but also credible, proportionate and constructive.
It also creates a useful learning opportunity for me. Reviewing these assessments allows me to recognise recurring coaching challenges, understand where the framework may need refining and identify themes that can inform our wider leadership development work.
In effect, we’re using AI to help evaluate the human coaching conversation, and human experience to evaluate the quality of the AI’s feedback.
I think that’s an important example of what responsible integration should look like.
Of course, this approach also raises questions about scale. Human verification takes time, and any organisation adopting a similar model needs to decide what level of review is appropriate for the importance and sensitivity of the feedback being produced. But efficiency should not come at the expense of trust, particularly when we’re evaluating people’s professional capabilities.
The objective is not to produce as many coaching reports as possible. It’s to produce feedback that managers can trust and use to become better coaches.
Supporting the salesperson and developing the manager
There are two important beneficiaries of this approach, and I think we need to give both equal attention.
For the salesperson, the benefit begins with the weekly conversation itself. They receive dedicated time with their manager to discuss opportunities, explore challenges and agree meaningful actions. They have access to somebody who understands their commercial environment, their customers and their individual development needs.
The conversation should help them leave better prepared for their next customer interaction, with greater clarity about what they need to achieve and why.
AI can support that process by helping make agreed actions, development priorities and learning points more visible and consistent. It can also provide additional guidance between meetings, helping the salesperson prepare and reflect without having to wait for the next scheduled conversation.
For the sales manager, the opportunity is equally significant. They receive feedback about the quality of their coaching that would otherwise be difficult to obtain consistently. They can identify recurring habits, recognise missed opportunities and establish specific areas for improvement.
Rather than simply being told that coaching is important, they have a practical way to develop their own capability as a coach.
And this is where the relationship between technology and human leadership becomes particularly valuable. The manager remains accountable for coaching, supporting and developing their people. AI helps them fulfil that responsibility more effectively, while the human verification of the feedback provides an additional safeguard around its quality.
In my view, that’s a far more compelling use of technology than attempting to automate the entire coaching relationship.
Slow at the start to go faster at the end
There’s a principle I’ve advocated throughout my sales career: slow at the start to go faster at the end.
We normally associate it with the sales process, particularly the importance of understanding the customer’s situation, establishing genuine need and qualifying opportunities properly before rushing towards proposals and negotiations.
But I think it applies equally to sales leadership.
Taking the time to help a salesperson understand why an opportunity isn’t progressing, develop a better approach to a customer conversation or challenge an assumption may not immediately improve this week’s numbers. But it should help develop better commercial judgement, stronger capability and more sustainable performance.
The challenge is that coaching requires consistency. It cannot be something we do occasionally when the forecast looks healthy or when the pressure of delivering results temporarily subsides.
It has to become part of the normal rhythm of sales leadership.
This is why I place so much importance on structured weekly conversations. They provide the discipline and continuity needed to connect development with the real work of selling.
AI can then help us improve the quality of those conversations, rather than simply measuring whether they happened.
And perhaps that’s the real opportunity we’ve been overlooking.
I've changed my mind, but not my belief in human coaching
I remain convinced that the relationship between a salesperson and their manager is fundamental to effective sales development. Trust, understanding, encouragement, challenge and the willingness to support somebody through difficult commercial situations cannot simply be delegated to a technology platform.
But I no longer see AI as an inevitable threat to that relationship.
Having better understood what is possible, and having seen how structured feedback can support real coaching conversations, I now see an opportunity to strengthen something I have long believed we need more of.
We can provide salespeople with more consistent guidance and support, while helping managers develop the coaching skills that should be central to their leadership responsibilities.
And we can do so through a process that combines human conversation, AI-supported analysis and experienced human verification, rather than assuming any one of those elements is sufficient on its own.
That doesn’t mean every AI coaching application will be effective, or that technology alone will resolve the wider problem of declining investment in people development. The quality of the framework, the coaching process, the human oversight and the leadership commitment will ultimately determine whether it delivers value.
But it does mean we have an opportunity to approach the debate differently.
Rather than asking whether AI will replace sales coaching, perhaps we should be asking how it can help us make coaching more consistent, more effective and more valuable to everyone involved.
Because the real challenge facing sales leadership may not be that artificial intelligence is becoming capable of coaching salespeople.
It may be that we’ve become so focused on managing results that we’ve neglected our responsibility to develop the people who deliver them.
And if AI can help us reverse that trend, I think that’s something worth embracing.
So here’s my question for sales leaders: if you listened to the last four weekly conversations between your managers and their salespeople, would you hear evidence of people being developed, or simply evidence of business being managed?
